INDEX

Explanations

someone or people related

The neuron activates on general references to people (e.g. “someone,” “other,” “person,” “them”).

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

of

-1.59

 Comunicação

-1.06

 harus

-1.05

全部

-1.05

难受

-1.03

to

-1.03

↵↵↵↵↵↵↵↵↵↵↵↵

-1.00

 faktisk

-1.00

瞅

-0.99

꒳

-0.98

POSITIVE LOGITS

 whom

1.75

 when

1.20

who

1.16

 keď

1.12

 cuando

1.09

 transferir

1.09

 wegen

1.08

whom

1.07

 しま

1.06

if

1.05

Activations Density 0.080%